Influence of Water Content on Glass Transition Temperature

The depression of the glass transition temperature of a wheat gluten film by water was studied by dynamic mechanical thermal analysis and differential...
9 downloads 0 Views 152KB Size
J. Agric. Food Chem. 1996, 44, 3474−3478

3474

Edible Wheat Gluten Film: Influence of Water Content on Glass Transition Temperature N. Gontard*,† and S. Ring‡ ENSIA-SIARC/CIRAD-SAR, B.P. 5098, 34032 Montpellier Cedex, France, and IFR/BBSR 2, Colney Lane, Norwich NR47 UA, United Kingdom

The depression of the glass transition temperature of a wheat gluten film by water was studied by dynamic mechanical thermal analysis and differential scanning calorimetry. Wheat gluten film behaved as a typical amorphous polymer, and the large plasticizing effect of water did not fit simply on the prediction of the Couchman-Karasz equation. Temperature and water mass fraction, at which large changes of film mechanical and barrier properties were previously observed, are well correlated with glass transition. Keywords: Edible film; wheat gluten; glass transition INTRODUCTION

Functional properties of edible or biodegrable films from food biopolymers are highly influenced by numerous parameters, such as film-forming technology, solvent characteristic, additives content, etc. (Guilbert, 1986; Kester and Fennema, 1986; Guilbert and Biquet, 1989). Because of their hydrophilicity, film properties are also highly dependent on usage conditions such as relative humidity and temperature (Kamper and Fennema, 1984; Biquet and Labuza, 1988; Rico Pena and Torres, 1991; Gontard et al., 1993). Water is the most ubiquitous and uncontrollable plasticizer. Until recently, this temperature and water content dependence was only interpreted in terms of water activity or disruptive water-polymer hydrogen bonding in a polymer hydrogen-bonded network. The structure-properties relationships of hydrated biopolymers could be better understood with the theories of glass transition used in polymer science; that is, in terms of critical variables of time, temperature, and diluent content (Slade, 1984; Levine and Slade, 1987, 1991; Slade et al., 1989). Glass transition of an amorphous or partially amorphous polymer is typically described as a transition from a brittle glass to a highly viscous or a liquid-like rubber (Levine and Slade, 1991; Allen, 1993). The glass transition separates two domains according to clear structural differences, and polymer properties are seriously modified when the temperature rises above the glass transition temperature (Tg). Theoretically, the addition of plasticizers such as water under isothermic conditions has the same effect as increased temperature on molecular mobility. Water, which is a low molecular weight component, increases free volume and thereby allows increased backbone chain segmental mobility. In spite of their wide variety of structure, reflecting the diversity of the aminoacids from which they are built, it is well established that most proteins undergo a glass transition. Plasticization by water affects the Tg of amorphous or partially amorphous proteins such as caseine (Kalichevsky et al., 1993), gelatin (Yannas, * Author to whom correspondence should be addressed (fax 33 67 61 70 55; e-mail [email protected]). † ENSIA-SARC/CIRAD-SAR. ‡ IFR/BBSR 2. S0021-8561(96)00230-0 CCC: $12.00

1972; Marshall and Petrie, 1980), collagen (Batzer and Kreibich, 1981), or elastin (Kakivaya and Hoeve, 1975; Scandola et al., 1981), resulting in a drop of Tg. A variety of techniques including differential scanning calorimetry (DSC), dynamical mechanical thermal analysis (DMTA), or pulsed nuclear magnetic resonance, have been used to show that wheat gluten is a highly amorphous multipolymer system that is water platicizable but not water soluble (Hoseney et al., 1986; Slade, 1984; Slade et al., 1989; Kalichevsky et al., 1992). In previous investigations (Gontard et al., 1991, 1992, 1993), an edible wheat gluten film was developed, and the effects of film-forming conditions, plasticizer content (glycerol), additives, relative humidity, and temperature on various film properties (water vapor permeability, mechanical, sorption properties) were studied. In the present study, the Tg of wheat gluten film is studied as a function of water content by DSC and DMTA to gain a better understanding and control of the plasticizing effect of water on the film. MATERIALS AND METHODS Wheat Gluten Film Preparation. A film-forming solution was prepared with gluten (Amylum Aquitaine, F 33000 Bordeaux), glycerol (95%, Merck, Darmstadt, Germany), acetic acid, ethanol (Aldrich Chemie, Steinhem, Germany), and water. The gluten concentration (7.5 g/100 mL solution), glycerol concentration (20 g/100 g of gluten), ethanol concentration (45 mL/100 mL solution), and pH of the solution (4.0 adjusted with acetic acid) were chosen on the basis of results from previous investigations (Gontard et al., 1992, 1993). All components were mixed vigorously at 40 °C, and the filmforming solution was then immediately poured on a plexiglassleveled surface and dispersed with a thin-layer chromatography (TLC) spreader (Braive, 4000 Lie`ge, Belgium) adjusted to 0.8 mm height. The film-forming solution was dried in a ventilated oven at 30 °C to evaporate volatile solvents (ethanol, acetic acid, and water). This procedure resulted in the formation of a transparent film with a constant controlled thickness (0.050 ( 0.003 mm) that was measured at several positions with a micrometer (Roch). Disks (for DSC measurements) or rectangles (for DMTA measurements) of film were equilibrated for 5 days at 25 °C in atmospheres with controlled relative humidities (using saturated salt solutions of known water activity ranging from 0.06 to 0.97 at 30 °C (Spiess and Wolf, 1987; Multon, 1984; Stamp et al., 1984). The moisture content of each film was evaluated with previously determined sorption isotherms (Gontard et al., 1993).

© 1996 American Chemical Society

Glass Transition of Gluten Film

Figure 1. DSC thermogram of wheat gluten film stored at 0-90% relative humidity (water mass fraction, 0.008-0.224). DSC Measurements. Ten milligrams of small film disks of known dry matter were piled up in open stainless-steel pans, which were designed to withstand high pressures and suppress the volatilization of water, and equilibrated at the desired relative humidity for 5 days. After equilibration, the pans were immediately and hermetically sealed. Calorimetric measurements were carried out with a Perkin Elmer DSC-4 calorimeter equipped with a thermal analysis data station. The temperature was calibrated daily with the melting points of Indium and cyclohexane II. An empty aluminium pan was used as an inert reference. The Tg was determined from the midpoint of the heat capacity change observed at a heating rate of 10 °C‚min-1 on at least triplicate samples, and verified from the second run after cooling to eliminate the eventual effects of sample history. Dynamic Mechanical Thermal Analysis (DMTA) Measurements. DMTA measurements were made with a Polymer Laboratories PL-DMTA mark I apparatus (Loughborough, U.K.). Sample rectangles of film (12.0 × 6 mm) of known dry matter were previously equilibrated at the desired relative humidity for 5 days. After equilibration, they were immediately coated with silicon grease to avoid water evaporation and subjected to the tensile mode at a frequency of 1 Hz while the temperature was increased from -20 to 160 °C at 2 °C‚min-1. Sample displacement gave the storage modulus (E′), the loss modulus (E′′), and tan δ () E′′/E′). The storage modulus is a measure of the stiffness or rigidity of the material and is calculated as the ratio of the stress to the applied strain. The loss modulus reflects the ability of a material to dissipate mechanical energy by converting it to heat through molecular motion. As the specific molecular groups undergo rotation at a given temperature, a peak is observed in the loss modulus data due to the absorption of the energy applied by the mechanical spectrometer. Tan δ is a useful index of material viscoelasticity (Urzendowski and Pechak, 1992). The Tg was defined as usual by the tan δ peak or the inflexion point of the decrease in E′ on at least triplicate samples. RESULTS AND DISCUSSION

The Tg of wheat gluten films was determined by DSC as well as by DMTA measurements. DSC measurements (Figure 1) show only a change of the baseline toward a higher value, indicating that gluten is a highly amorphous polymer undergoing a glass transition phenomenon. But the heat capacity change is relatively small (∼5 cal/g) and induced a 10-20% experimental error (compared with a 5-8% error for DMTA measurements). The Tg, which was taken in the curves as the inflection point, decreases with increasing water content.

J. Agric. Food Chem., Vol. 44, No. 11, 1996 3475

Figure 2. Typical DMTA plot for a wheat gluten film stored under 43% relative humidity (0.032 water mass fraction) showing tensile modulus (E′) and tan δ as a function of temperature.

Figure 3. DMTA tensile modulus (E′) as a function of temperature for wheat gluten film with 0.08 (0), 0.032 (O), 0.057 (4), 0.194 (2) and 0.292 (b) water mass fraction.

Glass transition phenomenon in gluten film was confirmed by DMTA. A typical DMTA graph showing E′ (elastic modulus) and tan δ as a function of temperature is shown in Figure 2. Wheat gluten film clearly displays the behavior of an amorphous polymer: with increasing temperature, this amorphous protein traverses the glassy region, the transition region (tan δ peak and modulus drop), and a rubbery plateau. The size of the tan δ peak that is thought to reflect the volume fraction of the material undergoing the transition (Wetton, 1986) is similar to the typical tan δ peak of other totally amorphous biopolymers such as elastin (Lillie and Gosline, 1993), amylopectin, gelatin, or native wheat gluten (Kalichevsky et al., 1993). In the glassy state, the storage modulus E′ displays a value of ∼109 Pa and is insensitive to moisture content changes (Figure 3). At the glass-to-rubber transition, a characteristic decrease of 103 Pa is observed. These results are consistent with those for native wheat gluten or other biopolymers, such as casein (Kalichevsky et al., 1993), starch, gelatin, or ovalbumin (Kalichevsky et al., 1993; Yannas, 1972). The absence of a flow region precludes the presence of intermolecular covalent crossbondings, such as disulfide bonds. Gluten has been described as a thermosetting polymer (Slade et al., 1989) and glutenin subunits (the higher molecular weight component of gluten) are linked by interchain disulfide bonding of cysteine residues which maintains high rubber-like modulus. Glycerol, which was used as plasticizer, appears to be miscible with gluten at the composition used because only one transition is observed. However, the broaden-

3476 J. Agric. Food Chem., Vol. 44, No. 11, 1996

Figure 4. Tg of wheat gluten film determined by DMTA E′ inflexion point (O) and tan δ peak (4), and by DSC (9) as a function of water content. The solid and dashed lines are the eq 2 theoretical curve obtained for K1 ) 2.1 (- - -) and K1 ) 4.0 (s). The crosses indicate water mass fraction and temperature at which wheat gluten film changes of mechanical and barrier properties occur, according to Gontard et al. (1994).

ing of the transition (a few tens of °C) could suggest a lack of miscibility or simply reflect the high heterogeneity of gluten proteins molecular weights ranging from 30 000 to millions (Graveland et al., 1985; Bietz and Wall, 1972). The Tg decreased when the water content of the film increased (Figure 3). It is interesting to note that, as observed with wheat gluten (Kalichesky et al., 1992) or gelatin (Yannas, 1972), increasing the water content decreases the rubbery modulus (high-temperature value of E′). It has been observed that E′ values above Tg are highly dependent on the crosslink density of a polymer; thus, the reduction of the rubbery modulus in the presence of a plasticizer implies reduced cross-linking. In this case, the reduction in E′ may be due to polymerpolymer hydrogen bonds being replaced by labile polymer-water hydrogen bonds (Kalichevsky et al., 1992). The inflexion point in the drop in E′ and tan δ peak are plotted again water content in Figure 4 and can be compared with DSC results. A good correspondence is obtained when DSC results are compared with the E′ inflexion point. As previously reported for wheat gluten (Kalichesky et al., 1992) or other biopolymers such as casein (Kalichesky et al., 1993), the tan δ peak falls above the E′ inflection point and the midpoint in the change in heat capacity observed by DSC. A simple comparison between mechanical and calorimetric determinations of Tg is not easy. Techniques are sensitive to different degrees of molecular mobility. Glass transition is not observed at a unique temperature, but is related to the frequency and nature of the measurement technique. The Tg measured by DMTA is dependent on frequency, and Tg measured by DSC depends on the heating rate used because it is a kinetically determined transition. However, the DMTA method is a more sensitive method and the glass transitions were easier to detect. Thus, DMTA results are very reproducible for high water contents (5% of experimental error), where glass transition occured at a low temperature avoiding water loss. DSC results could be more reliable

Gontard and Ring

at low water content because no water loss occurred. These two methods could thus be considered as complementary. On increasing water content, the value of the Tg drops, at first rapidly (5 °C/1% water for 0-0.1 water mass fraction) and then more slowly (3.5 °C/1% water for 0.1-0.3 water mass fraction), indicating a large plasticizing effect of water on wheat gluten film. The sensitivity of Tg to polar diluent such as water makes it very likely that this transformation involves the regions that contain large polar residues such as Glu, Asp, Arg, and Lys. This lowering of Tg is similar to the value observed for gelatin/glycerol system (5 °C/1% water; Yannas, 1972), starch (∼5 °C/1% water; Orford et al., 1989), elastin (15 °C/1% for low water content and 5 °C/1% water for 0.1-0.2 water mass fraction; Kakivaya and Hoeve, 1975), and other biopolymers (generally ∼510 °C/1% water; Slade and Levine, 1989). The depression of Tg by the addition of a diluent or plasticizer may be explained by a number of theoretical approaches such as free volume theories or classical thermodynamic theories proposed by Couchman and Karasz (1978). A thermodynamic consideration of the entropies of the components at the Tg of the mixtures leads to the following equation for a three-component mixture:

Tg (mixture) ) W1∆Cp1Tg1 + W2∆Cp2Tg2 + W3∆Cp3Tg3 (1) W1∆Cp1 + W2∆Cp2 + W3∆Cp3 In eq 1, the subscript 1 refers to water, the subscript 2 refers to wheat gluten, and the subscript 3 refers to glycerol, W is the weight fraction of each component, and ∆Cp is the change in heat capacity observed at Tg. This equation requires values for the ∆Cp of water that have been subject of considerable debate (Kalichevsky et al., 1992). Because there is one less constant and no need for water ∆Cp, the Gordon-Taylor equation (Gordon and Taylor, 1952) has been more easily and successfully used to fit Tg data of various biopolymers such as starch (Roos and Karel, 1991). But, this latter equation cannot be extented to more than bicomponent system. Thus, to obtain a better fit to the predicted Tg of mixture, the following equation, which is equivalent to eq 1 but avoids the use of water ∆Cp, was used:

Tg )

W1K1Tg1 + W2Tg2 + 0.2W2K2Tg3 W1K1 + W2 + 0.2W2K2

(2)

In eq 2, 0.2 ) W3/W2 (equal to the constant weigth ratio of glycerol/gluten), K1 ) ∆Cp1/∆Cp2, and K2 ) ∆Cp3/∆Cp2. In addition, K is a constant that is proportional to the plasticizing effect of diluent (water for subscript 1 or glycerol for subscript 3) on the polymer (wheat gluten with the subscript 2). For dry film, W1 ) 0, and eq 2 becomes

Tg(mixture) ) Tg2 + 0.2K2Tg3/1 + 0.2K2

(3)

Using Tg2 ) 435 K and Tg3 ) 180 K (from Kalichevsky et al., 1992), and Tg(mixture) ) 384 K (dry film Tg determined in the present study), K2 was calculated from eq 3. The K2 value thus obtained was 1.28, which is a lower value than the one calculated (2.26) with dry gluten and glycerol ∆Cp (respectively, ∆Cp2 ) 0.39 and ∆Cp3 ) 0.88 J‚g-1‚K-1) from Kalichevsky et al. (1992). This result means that wheat gluten proteins were

Glass Transition of Gluten Film

plasticized by glycerol but not as much as predicted. Similarly, but to a greater extent, observations have been made on the plasticizing effects of fructose on sodium caseinate (Kalichevsky et al., 1993), or of fructose, sucrose, and glucose on wheat gluten (Kalichevsky et al., 1992). It seems that the incomplete compatibility of sugar and proteins was responsible for this phenomenon. Glycerol appears to be more miscible with gluten than sugars, because of its lower molecular weight, but the compatibility is not complete and could explain the deviation from the Couchman-Karasz theory. The DSC and DMTA data were fitted with eq 2, with K1 as a variable. As shown in Figure 4, eq 2 could not be easily correlated with the experimental curve with a unique K1 that represents the ratio of water ∆Cp to wheat gluten ∆Cp. With K1 ) 4.0, a very good correlation was obtained for water mass fraction of